Meetings/2009-04-29/DemoNotes

Demo - http://sharedname.org/page/Meetings/2009-04-29/Demo

5 servers - one "master" - databases switched to Restsync. Uses Mysql synchronization

http://demo.sharedname.org/local - multiple 'a' records, so round robins servers

"/n/" - top level domain on purl.oclc.org

"/n/pmid/" - won't allow more specific parts, so html in the middle (will be fixed)

Demo showed different redirects to pubmed documents

Harold: Why not content negotiation? Alan: What are the constraints on content negotiation - every URI doesn't content negotiate because content negotiation is (a) evil, and (b) http specific.

NOTE: Spec has to make content negotiation responses explicit.

Olivier - What if I make an annotation and want to use this as an annotation. I don't care if it is asn, xml, etc? Alan: (Showed top level URL and "about" redirect)


Two names -

  1. no encoding the record without commitment to version or encoding. Returns redict (303) to "about record
  2. about record - typically subject headings, etc. What this record is about.

Olivier: - what if there are multiple sources of pubmed? Alan: Would have mirror and authoritative identifiers. Other types of record that talk about subject matter? If something like refseq that has versions, versioning information might be here as well.

Harold: What about translations from existing links to shared names? A: on the roadmap, but not now. Harold, Benjamin and ??? volunteered to investigate reverse requirements and technology.

Q: how about additions for local enhancements and annotation? A: local is fine, but if common servers have non-standard behavior, then preferences will be for particular servers, so will not be supported in published services

Q: What of provenance information? A: Might have multiple links to, say, different rdfizations of this, but will need provenance information.

Can have database records about a thing before its been ontologized. Strictly speaking, it is out of scope but ...


  • Example of how to curate and translate into html and RDF in a wiki (off of the page)
  • Templates are generated directly and loaded the wiki defined templates.
  • Can open in Protégé RDF editor as well (???)